[00940]

The PCM2707 chip in my DAC decided to stop working, after a few months of good work (they're more sensitive to plug-in/out issues than the good old pcm2702). Desoldering went bad and I lifted some tracks (and lost a second pcm2707 trying to wire it in anyway

) so replacing it looks like a lost cause. The PCB will however accept I2S inputs so I "just" have to find a solution to go from USB to I2S.
 
The options I've found so far (fortunately, I've still a pair of PCM2707 laying around) :
 
- Twisted Pear's USB receiver. Bare PCB is 14$ (+2.5$ shipping)
- Amb's gamma1. Bare PCB is 12$  (+ ? shipping)
- Hacked grubdac. Cheapest kit is 35$  (maybe less if I can persuade Bezaar to sell me a "reduced" kit)
 
 
Any comment is welcome. My heart is leaning towards the TP board right now.

 

[cobaltmute]

You want USB to I2S?
 
I could throw you a quick board based on the grubDAC, ripping off the WM8524 and putting in a set of terminal blocks.

 

[00940]

What I had in mind was to simply solder a shielded cable to the different vias connected to the I2S lines on my dac (the shield being connected to ground) and do the same to the grubdac board. It should be quite easy and a stock pcb would be no problem. Vias don't lift easily either 

 

[holland]

are you going to be able to salvage other parts?  the grubdac PCB is $4.00 + shipping.
 
GrubDAC PCB

 

[00940]

Oopsie, I hadn't seen that Bezaar was selling bare boards.
 
Sadly there aren't many parts I could salvage (different regs, parts pinout issue, etc). Getting the kit anyway might be a sound decision as sourcing parts from somewhere else would generate new shipping costs and so on.

 

[00940]

I just checked what I was saying... getting the parts I need to populate the grubdac board (the tps79333, the clock, the smd caps, ferrites and resistors) would cost around 25€ from either RS or Farnell (you have to buy smd by 50 and shipping isn't cheap).
 
25€ is 35$, the price of Beezar's kit. I might just get it; that'll give me an extra pcm2707 to play with if I mess up again.

 

[nullstring]

FYI, cheap and flawless USB to I2S is a problem diyaudio folks have been trying to solve for a while now without success
http://www.diyaudio.com/forums/digital-source/167392-ultimate-usb-i2s-interface.html
 
Finally, it looks like we have a project called audio-widget that will solve it.
https://groups.google.com/group/audio-widget?pli=1
 
Still in the prototyping stages though.
 
 
Also, DDDAC sells a USB to I2S module for 49 euros.
http://www.dddac.de/
Just another option.

 

[00940]

Yep, it isn't such an easy problem. A pcm2707 level of performance is ok for me at this point. I've got a SRC4192 following the usb receiver and my music collection is pretty much coming only from CDs. So 44.1/16 is fine too.
 
I'm going to bookmark that audio-widget thing though and follow a bit how it goes on. Thanks for the link.
 
Yes, I know the DDDAC module. But it costs twice as much as the grubdac (60€ shipped vs about 30€) and offers me no additionnal features I'd want.

 

[00940]

I finally decided to get a board of my own done at Olimex. I need to order pcbs for other projects anyway so I'll add it to a panel (it's only 33mm/25mm).

The board is really stripped down:
- I2S output only,
- bus powered (less worry about connecting, disconnecting the cable),
- no additional regulator (if it's good enough for the Twisted Pear board, it's good enough for me),
- USB input and I2S output by 0.1" terminals.
- Audioengr once suggested to add 47pf caps after the 22R resistors on the D+,D- lines, so I made room for such caps.
 
It only uses common parts: 1206 caps and resistors, hc49 oscillator in order to avoid ordering from big distributors as, in Europe at least, it's not a very pleasant experience. DRC is set at 10mil. The ground pad at the bottom right is to ground the oscillator case.
 
I'm still designing the other pcbs going onto the panel, so please feel free to suggest improvements.

Schematic (beware: pins with overlapping gnd symbol will be cut)
 


Full pcb
 


Top only
 

 

[forsakenrider]

any updates on this? Im looking for something to do the same
 

 

[00940]

I received a pair of boards from Olimex today (along with a I/V pcb, a tpa6120 preamp pcb and a stand alone soha II buffer). They seem fine but I had no courage to start soldering smd today so I played with the buffer rather than with my DAC.

 
They're extremely small. I have one regret though: I should have provided for slightly bigger holes, to allow an easy fit of terminals. As is, it's a very tight fit.

 

[Nada]

Thus looks really interesting.
Can someone explain to a DIY noob, what controls the timimg for the I2S to keep jitter down?

 

[00940]

The pcm2707 is an adaptive usb receiver. John Swenson made a good explanation on audio asylum ( http://www.audioasylum.com/forums/pcaudio/messages/7719.html ):
 
Quote:

Adaptive: in this mode the clock comes from a separate clock generator (usually implemented as a PLL referenced by a crystal oscillator) that can have its frequency adjusted in small increments over a wide range. A control circuit (either hardware or firmware running on an embedded processor) measures the average rate of the DATA coming over the bus and adjusts the clock to match that. Since the clock is not directly derived from a bus signal it is far less sensitive to bus jitter than synchronous mode, but what is going on on the bus still can effect it. Its still generated by a PLL that takes its control from the circuits that see the jitter on the bus. Its a lot better than synchronous mode, but still not perfect by a long shot. This is the mode that MOST USB audio devices use today.

 
The jitter of the pcm2707 isn't that low. Some have measured it up to 2, 3ns in barebone implementations such as the one here. If you use high quality power supplies, dedicated clocks and so on, you can do better but it's still not the best usb receiver ever, it's very basic.
 
However, in this particular case, I'll be using it before an SRC4192 asynchronous sample rate converter (upsampler chip). This particular chip has the specificity to actually supress incoming jitter, as long as it is under a few dozens of ns (it was pointed out on diyaudio by Bruno Putzeys a long time ago). It isn't perfect, as the upsampling process produces some artifacts but the thd figures are so low that it is pretty negligible in my opinion and according to my listening experiences.

 

[00940]

Update: I soldered one and quickly connected it to the laptop. It seems fine and is correctly recognized by the computer.
 
However, as the DAC it'll go in is waiting to be re-cased, it'll be a little while before I can tell you if it actually works.

 

[forsakenrider]

Did you say you had "a couple" of the boards made? If so I would be interested in one!